What the data can't account for is who you lost and what your depth looked like.Interesting stuff here. 538 did an analysis of what losing starters means to a team's win total. Findings relevant to us:
https://fivethirtyeight.com/features/michigans-lineup-was-gutted-how-much-will-it-matter/
On offense:
I'm still figuring it all out, but looks like we can expect to be down 1-2 wins against an average schedule compared to your average team.
I mean, if you lose a starting QB and someone else loses a starting OG it is not the same thing. And if you lose a starting QB but your depth is a guy who started for you in fewer than 4 games (therefore a non-starter by the definition these sites use), playing well, that's different than replacing with unproven depth.
I look at CU's roster and I kind of shrug at the numbers when they're telling me that guys like Oliver, Montez, Huckins and McCartney aren't returning starters.